Papers by Liane Lewin-Eytan
Generating Q&A Benchmarks for RAG Evaluation in Enterprise Settings (2025.acl-industry)
Copied to clipboard
| Challenge: | DataMorgana generates synthetic Q&A benchmarks tailored to RAG applications . lexical, syntactic, and semantic diversity of generated benchmarks exceeds existing tools . |
| Approach: | They propose a tool for generating synthetic Q&A benchmarks tailored to RAG applications in enterprise settings. |
| Outcome: | The proposed tool surpasses existing tools in terms of lexical, syntactic, and semantic diversity while maintaining high quality. |